GSM8K
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GSM8K
103.69Tokens
30
GSM8K
0.93Accuracy
30
GSM8K sampled subset n=200 (test)
95.5Accuracy
30
GSM8K Numerical +4 (test)
88Accuracy
30
GSM8K
77.13Accuracy
29
GSM8K (test)
85Accuracy
29
GSM8K
93.9Accuracy
29
GSM8K v1.1 (test)
81.7Accuracy
28
GSM8K
89.47FA
28
GSM8K
73.39Retention
28
GSM8K (test)
94.04Mean@4
28
GSM8K
81.96Accuracy
28
GSM8K official train (dev)
88.3Accuracy
28
GSM8K ID
66.9Greedy Accuracy
27
GSM8K
86.4Accuracy
27
GSM8K
0.789AUROC
27
GSM8K (test)
94.47Pass@4 Score
27
GSM8K
96.7Accuracy
27
GSM8K
92Detection Accuracy
27
GSM8K 1,000 samples (val)
0.668AUROC
27
GSM8k
2,087Throughput (tok/s)
27
GSM8K
90.22Solve Rate
27
GSM8k
95.2Pass@1
27
GSM8K
71.87Accuracy
26
GSM8K
3.88Speed (×)
26